Why ss2tf changes the dynamics of model

Hi all,

I am working on converting a state space model to a Partial Fraction Expansion. My idea is first convert the state space to transfer function by 'tf2ss', then obtain the Partial Fraction Expansion by command 'residue'.

However, when I tried converting my state space model with 33 dimensions to a transfer function, the obtained transfer function gives different zero-pole plot, as well as bode plots. Anyone has any insight about why they are not consistent?

Or even better, anyone know how to obtain a Partial Fraction Expansion from state space model?

Thank you very much!

Answer 1

[Aykut Satici]

Anytime you convert from one model type to another in MATLAB, there is some accuracy lost due to the numerical accuracy of the computations. You can read more about these in the following link:

http://www.mathworks.com/help/control/ug/conversion-between-model-types.html

For instance, if you do not have a minimal realization of your model in state space, then the conversion may lead to two systems whose behavior is different. Consider finding a minimal realization of your system either by hand or by using MATLAB's "minreal" function. This function eliminates uncontrollable or unobservable states in the state-space models leading to a minimal realization of your system. You can read more about this function here:

http://www.mathworks.com/help/control/ref/minreal.html

[pippo]

Thanks for the reply. You are correct about it. My issue has beee solved by avoiding the ss2tf command.